EVA-0: Test-Time Model Evolution with Only Two Forward Passes per Sample

• URL: http://arxiv.org/abs/2605.18867v1
• Date: Fri, 15 May 2026 09:26:02 GMT
• Title: EVA-0: Test-Time Model Evolution with Only Two Forward Passes per Sample
• Abstract summary: We study test-time model evolution under a strict two-forward budget.<n>We reveal three key obstacles in zeroth-order test-time optimization.<n>We propose EVA-0, a minimal zeroth-order adaptation framework.
• Score: 55.284129005947484
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Test-time model evolution offers a promising way for deployed models to improve from unlabeled test-time experience, yet most existing methods depend on backpropagation (BP), which incurs substantial memory overhead and makes them difficult to deploy on edge devices, quantized models, specialized accelerators, or black-box models. In this work, we study test-time model evolution under a strict two-forward budget, a setting that pushes adaptation toward highly efficient real-world deployment. We reveal three key obstacles in zeroth-order test-time optimization: susceptibility to shortcut solutions, uncontrolled weight drift, and ineffective update direction estimation. To overcome them, we propose EVA-0, a minimal zeroth-order adaptation framework that: 1) keeps the loss scale-invariant to prevent shortcut solutions; 2) devises an anchor-guided optimization strategy to alleviate weight drift; 3) uses sample-wise symmetric two-sided perturbation for update direction estimation and inference. EVA-0 requires no BP and performs both inference and adaptation within only two forward passes per sample. Results on ImageNet-C & ViT-Base show that EVA-0 outperforms both BP-based DeYO and BP-free FOA, while achieving a 14x speed-up over FOA. Code will be released.

Related papers

• Continual Visual Anomaly Detection on the Edge: Benchmark and Efficient Solutions [11.262875405161417]
  Visual Anomaly Detection (VAD) is a critical task for many applications including industrial inspection and healthcare.<n>Our benchmark provides guidance for the selection of the optimal backbone and VAD method under joint efficiency and adaptability constraints.<n>We propose the first comprehensive benchmark for VAD on the edge in the continual learning scenario, evaluating seven VAD models across three lightweight backbone architectures.
  arXiv  Detail & Related papers  (2026-04-07T20:19:34Z)
• ZOTTA: Test-Time Adaptation with Gradient-Free Zeroth-Order Optimization [29.670784134543027]
  Test-time adaptation aims to improve model robustness under distribution shifts.<n>Most existing methods rely on backpropagation (BP), which is computationally costly and incompatible with non-differentiable models.<n>We propose ZOTTA, a fully BP-free TTA framework that performs efficient adaptation using only forward passes via Zeroth-Order Optimization.
  arXiv  Detail & Related papers  (2026-03-15T07:21:01Z)
• Enhancing Test Time Adaptation with Few-shot Guidance [62.49199492255226]
  Deep neural networks often encounter significant performance drops while facing with domain shifts between training (source) and test (target) data.<n>Test Time Adaptation (TTA) methods have been proposed to adapt pre-trained source model to handle out-of-distribution streaming target data.<n>We develop Few-Shot Test Time Adaptation (FS-TTA), a novel and practical setting that utilizes a few-shot support set on top of TTA.
  arXiv  Detail & Related papers  (2024-09-02T15:50:48Z)
• DPO: Dual-Perturbation Optimization for Test-time Adaptation in 3D Object Detection [34.04061546178302]
  Training 3D detectors in the real world often yields unsatisfactory performance when the distribution of the test data significantly deviates from the training data. We propose textbfdual-perturbation optimization (DPO) for textbfunderlineTest-underlinetime underlineAdaptation in underline3D. We minimize the sharpness to cultivate a flat loss landscape to ensure model resiliency to minor data variations.
  arXiv  Detail & Related papers  (2024-06-19T23:46:08Z)
• Test-Time Model Adaptation with Only Forward Passes [68.11784295706995]
  Test-time adaptation has proven effective in adapting a given trained model to unseen test samples with potential distribution shifts. We propose a test-time Forward-Optimization Adaptation (FOA) method. FOA runs on quantized 8-bit ViT, outperforms gradient-based TENT on full-precision 32-bit ViT, and achieves an up to 24-fold memory reduction on ImageNet-C.
  arXiv  Detail & Related papers  (2024-04-02T05:34:33Z)
• Towards Continual Learning Desiderata via HSIC-Bottleneck Orthogonalization and Equiangular Embedding [55.107555305760954]
  We propose a conceptually simple yet effective method that attributes forgetting to layer-wise parameter overwriting and the resulting decision boundary distortion. Our method achieves competitive accuracy performance, even with absolute superiority of zero exemplar buffer and 1.02x the base model.
  arXiv  Detail & Related papers  (2024-01-17T09:01:29Z)
• REALM: Robust Entropy Adaptive Loss Minimization for Improved Single-Sample Test-Time Adaptation [5.749155230209001]
  Fully-test-time adaptation (F-TTA) can mitigate performance loss due to distribution shifts between train and test data. We present a general framework for improving robustness of F-TTA to noisy samples, inspired by self-paced learning and robust loss functions.
  arXiv  Detail & Related papers  (2023-09-07T18:44:58Z)
• Sample-Efficient Optimisation with Probabilistic Transformer Surrogates [66.98962321504085]
  This paper investigates the feasibility of employing state-of-the-art probabilistic transformers in Bayesian optimisation. We observe two drawbacks stemming from their training procedure and loss definition, hindering their direct deployment as proxies in black-box optimisation. We introduce two components: 1) a BO-tailored training prior supporting non-uniformly distributed points, and 2) a novel approximate posterior regulariser trading-off accuracy and input sensitivity to filter favourable stationary points for improved predictive performance.
  arXiv  Detail & Related papers  (2022-05-27T11:13:17Z)
• Efficient Test-Time Model Adaptation without Forgetting [60.36499845014649]
  Test-time adaptation seeks to tackle potential distribution shifts between training and testing data. We propose an active sample selection criterion to identify reliable and non-redundant samples. We also introduce a Fisher regularizer to constrain important model parameters from drastic changes.
  arXiv  Detail & Related papers  (2022-04-06T06:39:40Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.